Sliding mode output tracking control based on a fuzzy clustering model

A sliding mode output tracking control method based on a fuzzy clustering model has been proposed in this paper. This technique describes the controlled system by rule-based fuzzy models. Parameters (antecedent membership functions, consequent parameters, etc.) in this structure can be fine-tuned with input-output data. By means of this technique, the complicated system can be decomposed into a number of simpler local models with a characteristic of smooth transitions between these sub-models by means of overlap membership functions. Then the implicit membership functions have been transformed into explicit functions of the antecedent variables with neural networks. Thus the control actions of a discrete sliding mode controller can be designed. Simulative results demonstrate the effectiveness of this method.